What is it about?
With the growing emphasis on ESG (Environmental, Social, and Governance) issues, the mandatory disclosure of ESG reports is on the horizon. However, due to the lack of a regulatory framework and a unified international ESG evaluation, the phenomenon of greenwashing in corporate ESG reporting is prevalent. We collected social responsibility reports and actual ESG performance data from A-share companies from 2011 to 2021 and innovatively employed text mining techniques to quantitatively investigate the extent of greenwashing in ESG reports. Our study initially utilized the Word2Vec method, combined with Skip-gram and Continuous Bag of Words models to train word vectors, and built an ESG lexicon using seed words. ESG reports is subsequently segmented based on a defined sentence splitting function and TF-IDF algorithm is employed to extract keywords. By matching the keywords with the ESG lexicon, we precisely extracted the annual ESG discourse for each company and conducted sentiment analysis to derive a greenwashing score. Heterogeneity analysis reveals that firm ownership has no significant impact on the level of greenwashing, yet the industry and region in which the enterprise operates considerably influence the greenwashing level. This study holds implications for enhancing the quality of ESG reporting and optimizing investment decisions.
Featured Image
Read the Original
This page is a summary of: Detection of greenwashing in ESG reports of Chinese listed companies based on Word2vec and TF-IDF, March 2024, ACM (Association for Computing Machinery),
DOI: 10.1145/3655497.3655513.
You can read the full text:
Contributors
The following have contributed to this page